v0.135.0
Summary
The SDK introduces the execute_command_sync method for the workspace sandbox service and adds new metadata fields for machine learning schedules, features, and publish specifications. Breaking changes remove the first_token_timeout field from model routing configurations and drop request-tag parameters and enum values from catalog rate limits.
Summary generated by brickster.ai. For the full changelog and any code/binary attachments, follow the GitHub link above.
More from databricks/databricks-sdk-py
This release adds workspace services for AI Functions and Sandboxes, alongside new Postgres snapshot management and feature engineering backfill methods. Breaking changes make source schemas and tables optional in the Pipelines API and remove multiple fields across Catalog and ML services.
Add enabled_telemetry_features field for databricks.sdk.service.serving.TelemetryConfig.
The SDK adds git integration and auto-deploy capabilities for Databricks Apps, along with custom UDF support for ML functions. Pipeline responses now include effective serverless compute IDs for better visibility into resource allocation.
The action_type and status fields in the bundledeployments.Operation model are no longer required. Additionally, the create_operation method has been removed from the workspace-level bundle_deployments service.
The SDK now supports LinkedIn Ads and Marketo as pipeline connectors and expands PostgreSQL synced table support with halfvec and varchar types. The list_workspace_assignments methods now require pagination handling, which is a breaking change for existing code calling these APIs.
This release adds pagination to four IAM methods in AccountIamV2API and WorkspaceIamV2API (list_direct_group_members and list_workspace_assignment_details), which is a breaking change requiring code updates if you use these methods. Several new fields are added across services including agent_type for Genie conversations, group_id for direct group members, effective_serverless_compute_id for job runs, and pipeline_channel for Postgres pipelines.